Processor name

Hi,

Thanks to all who have been to the axon conference last week. I really enjoyed taking with many of you.

If think “hope” I have a simple questions.

We are optimizing thread/segments/batchsize and now the interceptors.

For the events and the commands the documentation does nor work out of the box.
I made it work nicely now but I think you need to improve documentation here (the documentation only works for interceptor if you not using sagas).

But my main question is. How can I set the TrackingProcessorName (which is displayed in the dashboard).

Currently we have the full patch name there. Which is long and does not say a lot…

Best, Michael

Hi Michael,

I need to fully agree with your first point.
It was a tremendous pleasure to talk with you and others at the AxonIQ conference last Friday!

Added, I must also agree with your documentation point.
We still have lots to improve to make traversal clearer, and definitely the configuration bits need to be expanded upon.
If you have style/ordering suggestions how we would tackle this in the best fashion possible, I am definitely all ears.

Now, to your actual question, which indeed is quite simple.
The easiest approach to rename the Event Processor from utilizing the package name, is by specifying the @ProcessingGroup annotation on the Event Handling Components.
Doing so will firstly group Event Handling Components under the same group, but would also tell the Event Processor to use it as it’s name (if nothing more specific is defined through the EventProcessingConfigurer).

There’s a small snippet in the Reference Guide which does show this annotation on class level, the replaying section to be exact.
Granted though, this is indeed hard to find and not overly clear…my apologies for that.

Hope this helps you out Michael!

Cheers,
Steven

Do we need to update the database when we change this or does axon take care?

Best, Michael

Hi Michael,

The framework does not have the smarts to deduce if you have renamed a Tracking Event Processor, sadly.
Granted, I am not sure how we would tackle such an API just yet.

However, this indeed means you will have to perform a change on your token_entry table to adjust the names.
Or, if the Event Handling Components and their processors are configured right, just having it “start from scratch” with a new token shouldn’t hurt.

With “configured right”, I mean that you’ve specified which methods can be replayed and which cannot (with the @DisallowReplay annotation for example) and that some Tracking Event Processor which introduce side effects (like sending emails as a response to handling an event) always receive a token at the head of the stream.

Hope this helps Micheal!

Cheers,
Steven